3424
Biochemistry 1994, 33, 3424-343 1
Kinetic Characteristics of Phosphofructokinase from Bacillus stearothermophilus: MgATP Nonallosterically Inhibits the Enzyme? Malcolm Byrnes, Xiaomao Zhu, Ezzat S. Younathan,' and Simon H. Chang' Department of Biochemistry, Louisiana State University, Baton Rouge, Louisiana 70803 Received September 28, 1993; Revised Manuscript Received January 6, 1994" ABSTRACT: The kinetic mechanism of phosphofructokinase from Bacillus sterothermophilus has been
investigated using steady-state measurements. The double-reciprocal patterns observed for initial velocity, product inhibition, and mixed alternate substrate studies of the reverse reaction establish that the mechanism involves rapid-equilibrium random binding of substrates and the formation of an abortive complex composed of enzyme, MgADP, and fructose 6-phosphate (E-MgADP-Fru-6P). Initial velocity patterns for the forward reaction show significant nonlinearity and resemble those seen for competitive substrate (MgATP) inhibition of an enzyme that obeys a random mechanism. A mutant BsPFK enzyme (GV212) was used to show that the inhibition is not due to MgATP binding in the effector site. Product and dead-end inhibition studies of the forward reaction are consistent with a random mechanism, after taking into account the effects of substrate inhibition by MgATP. Initial velocity measurements at low MgATP concentration show that the binding of MgATP is not a rapid-equilibrium process; Le., the rate of catalysis is faster than the rate of substrate binding. It is concluded that the kinetic mechanism of the forward reaction is sequential random, with the rate of MgATP binding slower than the catalytic rate. A model is presented that incorporates these results and proposes that substrate binding proceeds through two alternative pathways, one of which is kinetically disfavored. The observed MgATP substrate inhibition arises from both reaction flux through the disfavored pathway and, to some extent, abortive binding of MgATP in the Fru-6P site.
Phosphofructokinase(PFK,' EC 2.7.1.1 1) catalyzes the first committed step of glycolysis, the transfer of a terminal phosphate from ATP to /3-D-fructose 6-phosphate (Fru-6P) to form ADP and @-fructose1,6-bisphosphate (Fru-l,6BP): MgATP
+ Fru-6P
-
MgADP
+ Fru- 1,6BP
The allosteric PFK from the thermophilic bacterium Bacillus stearothermophilus, like its counterpart from Escherichia coli, is a tetramer of identical subunits, each composed of 319 amino acids (320 for E . coli PFK). Fifty-five percent of the amino acids are identical between the two bacterial enzymes (French & Chang, 1987). Both B. stearothermophilus PFK (BsPFK) and E . coli PFK (EcPFK) are allosterically inhibited by phosphoenolpyruvate (PEP). This inhibition is reversed by the activators adenosine diphosphate (ADP) and guanosine diphosphate (GDP) (Valdezetal., 1989; Blangyetal., 1968). A comparison between the crystal structures of BsPFK and EcPFK (Evans et al., 1981; Shirakihara & Evans, 1988) reveals striking structural similarity. The two enzymes have all the same secondary structural elements, and their subunit a-carbon traces are nearly superimposable. In both enzymes, the subunit is divided into a large and a small domain, with the active site located within the subunit in a cleft between the two domains. The active-site residues that bind substrates are either the same or similar between the two enzymes. Supported by National Institutes of Health Grant DK31676.
* Authors to whom correspondence should be addressed.
Abstract published in Advance ACS Abstracts, March 1, 1994. Abbreviations: PFK, phosphofructokinase; BsPFK and EcPFK, the phosphofructokinasesfrom Bacillus stearothermophilus and Escherichia coli, respectively; GV212, mutant B. stearothermophilus phosphofructokinase with glycine 212 mutated to valine; bspfk, the B. stearothermophilus phosphofructokinase gene; Fru-6P, fructose 6-phosphate; Fru1,6BP, fructose 1,6-bisphosphate; PEP, phosphoenolpyruvate; Ara-SP, arabinose 5-phosphate; AMPPCP, &y-methyleneadenosine 5'-triphosphate; AMPPNP, adenylyl imidodiphosphate; SDS, sodium dodecyl sulfate. @
0006-2960/94/0433-3424$04.50/0
Despite the high degree of structural similarity between BsPFK and EcPFK, there is a significant kinetic difference between them. Steady-state initial velocity studies reveal that whereas saturation of EcPFK by Fru-6P is highly cooperative in the presence of saturating MgATP (Blangy et al., 1968), saturation of BsPFK by Fru-6P shows little or no cooperativity under the same conditions (Valdez et al., 1989). Fructose 6-phosphate saturation of BsPFK is cooperative, however, in the presence of PEP (Valdez et al., 1989). Thus, BsPFK resembles the PFK from the extreme thermophilic bacterium Flavobacterium thermophilum (Yoshida, 1972), which likewise displays a hyperbolic Fru-6P saturation profile that become sigmoidal in the presence of PEP. Schirmer and Evans (1990) have proposed a structural model for the allosteric transition of BsPFK based on X-ray diffraction data. Although the structure of BsPFK has been extensively investigated, its kinetic characteristics have remained largely unstudied. For example, although the amino acid residues involved in binding the substrates MgATP and Fru-6P have been known for some time (Evans et al., 1981), the kinetic mechanism has not been studied. Several observations suggest that there are differences between the active-site regions of BsPFK and EcPFK; (i) the active-site structure of BsPFK is moreopen than that of EcPFK (Shirakihara & Evans, 1988), (ii) significant binding of GDP occurs in the active site of BsPFK (Valdez et al., 1989) but not in the active site of EcPFK (Blangy et al., 1968), and (iii) thelarge (ATP-binding) domain of EcPFK moves as the enzyme changes conformation from a "closed" to an "open" subunit structure (Shirakihara & Evans, 1988), whereas the large domain of BsPFK remains essentially rigid during the allosteric transition of the enzyme (Schirmer & Evans, 1990). These observations have stimulated our interest in studying the kinetic characteristics of the BsPFK-catalyzed reaction in both the forward and reverse directions. Initial velocity, product inhibition, and mixed alternate substrate studies of the reverse reaction indicate that the kinetic mechanism is 0 1994 American Chemical Society
8. stearothermophilus PFK Kinetics sequential random. Product and dead-end inhibition studies of the forward reaction corroborate this result. However, initial velocity studies of the forward reaction indicate that MgATP is a nonallosteric substrate inhibitor, and the binding of MgATP is a non-rapid equilibrium process. On the basis of these results, it is proposed that substrate binding proceeds via two alternative steady-state pathways, with one pathway kinetically favored over the other. Substrate inhibition by MgATP is proposed to result from both abortive binding of MgATP in the Fru-6P site and reaction flux through the disfavored pathway. MATERIALS AND METHODS Enzymes and Chemicals. NADH, NADP+, Fru-6P, ATP, Fru-l,6BP, ADP, PEP, Ara-5P, AMPPCP, and the auxiliary enzymes for the coupled PFK activity assays were all obtained from Sigma Chemical Co. (St. Louis, MO). The Cibacron Blue 3GA-agarose (Type 3000-CL-L) resin was also obtained from Sigma. Restriction endonucleases used for cloning were obtained from either New England Biolabs (Beverly, MA) or U S . Biochemical (Cleveland, OH). Expression and PurificationofPFK. BsPFK was expressed in PFK-deficient E. coli cells (DF1020 cells) from a recombinant plasmid constructed by cloning the bspfk gene (French & Chang, 1987) into the EcoRI/HindIII sitesof pUC18. The enzyme was subsequently purified by a three-step procedure (Valdez et al., 1989) involving (1) sonication, (2) heat treatment at 70 OC, and (3) chromatography on a Cibacron Blue 3GA affinity column. The purified enzyme preparation had a specific activity of 160 units/mg. Its purity was demonstrated by the presence of a single band on an SDSpolyacrylamide gel stained with Coomassie Brilliant Blue dye. The purified enzyme solution was dialyzed at 4 "C in 100 mM Tris-HC1, pH 7.4, containing 1 mM dithiothreitol and 50% glycerol, and stored at -20 OC. Protein concentration was determined using the Bio-Rad (Richmond, CA) protein assay. Site-Directed Mutagenesis and Expression and Purification of Mutant BsPFK. The site-directed mutagenesis procedure to mutate the bspfk gene, as well as the procedures for expression and purification of the mutant enzyme, are to be described separately (Zhu et al., manuscript in preparation). Kinetic Assays. The initial velocity of the PFK-catalyzed reaction in the forward direction was measured at 30 OC2 in 100 mM Tris-HC1, pH 8.2, containing 10 mM MgC12 and 5 mM NH4Cl by coupling the production of either Fru- 1,6BP or ADP to the oxidation of NADH (0.20 mM). The Fru1,6BP-coupled assay utilized the auxiliary enzymes aldolase (20 pg/mL), triosephosphate isomerase (10 pg/mL), and cy-glycerophosphate dehydrogenase (10 pg/mL), while the ADP-coupled assay utilized pyruvate kinase (10 pg/mL), phosphoenolpyruvate (PEP, 200 pM) and lactate dehydrogenase (10 pg/mL). The ADP-coupled assay system could not be used for Fru-6P saturation studies since PEP is an allosteric inhibitor with respect to Fru-6P; however, PEP at thisconcentration (200pM) does not alter MgATP saturation behavior. The ATP-regenerating system that utilizes creatine kinase and creatine phosphate, which is required for the EcPFK activity assay, was found to be unnecessary since ADP has little or no activating effect on BsPFK (in the absence of PEP). In all forward reaction assays, the free Mg2+ con-
* The saturation curves with respect to substrates Fru-6Pand MgATP obtained at 30 O C were similar to those obtained at 60 O C , except that the k,, was lower at 30 OC. Fru-6P saturation was noncooperative at both temperatures, and substrate inhibition was apparent with high [MgATP] at both temperatures.
Biochemistry, Vol. 33, No. 11, 1994 3425 centration was kept 5-10 mM in excess of the ATP concentration to avoid inhibition by free ATP. Typically, assays were initiated by the addition of 0.10 pg of PFK. Since the BsPFK assay generally involved an initial nonlinear phase, a 1- or 2-min lag period was allowed prior to recording the Abs340 change over time, which was done for at least 1 min. A thermostated Hitachi UV-2000 spectrophotometer was used for these measurements. Duplicate assays were run for each data point. Initial velocities are expressed in units of micromoles of product formed per minute. Initial velocity measurements in the reverse direction were similar to those in the forward direction, except that the production of either Fru-6P or ATP was coupled to the reduction of NADP+ (0.20 mM). Assays were run at pH 8.2, and Mg2+concentration was kept 9-10 mM in excess of the ADP concentration. For the Fru-6P-coupled assay, the auxiliary enzymes phosphoglucoisomerase ( 10 pg/mL) and glucose-6-phosphate dehydrogenase (10 pg/mL) were used, whereas for the ATP-coupled assay, hexokinase (20 pg/mL), glucose (3 mM), and glucose-6-phosphate dehydrogenase (10 pg/mL) were used. Typically, assays were initiated with the addition of 2.0 pg of PFK. Treatment of Kinetic Data. Initial velocity data for the forward reaction were fit to either the Michaelis-Menten equation (eq l), the Hill equation (eq 2), the initial velocity equation for a rapid-equilibrium sequential mechanism (eq 3; Cleland, 1963), or the initial velocity equation for a sequential random mechanism assuming steady-state (non-rapid-equilibrium) conditions (eq 4; Ferdinand, 1966). In eqs 1-7, [A] and [B] represent the concentrations of substrates A and B, and V, is the maximum velocity. In eq 1, K , is the Michaelis constant, whereas in eq 2, [ A ] * p is the concentration of substrate at half-saturation and n is the Hill coefficient. In eq 3, Kia is the equilibrium constant for dissociation of substrate A from the binary complex EA (E represents enzyme). There is an analogous constant, Kib, for dissociation of B from EB. Ka and Kb are equilibrium constants for dissociation of A and B, respectively, from the ternary complex EAB. In eq 4, the terms i, j, k, 1, and m are complex functions of [B] and the rate constants for thevarious steps in Scheme 1 (see Discussion section). As such, i-m have no physical significance.
i[A]' u=
k
+ j[A]
+ /[A]'+
m[A]
(4)
Initial velocity data for product inhibition, dead-end inhibition, and alternate substrate studies were first plotted graphically as double-reciprocal plots. On the basis of these primary plots, the inhibition patterns were identified, and each data set was fit to the equation for either linear competitive (eq 5), linear noncompetitive (eq 6 ) , or linear uncompetitive (eq 7) inhibition (Cleland, 1979). In eqs 5-7, the constant Kis the equilibrium constant for dissociation ofvaried substrate A from the ternary complex EAB, whereas Ki, and Kii are equilibrium constants for dissociation of inhibitor I from its
3426 Biochemistry, Vol. 33, NO. 11, 1994
Byrnes et al.
inhibitory complex, as determined from slope and intercept replots, respectively. All curve-fitting to eqs 1-7 was peru=
Vmax[Al K(1 + [Il/Kis) + [AI
(5)
1.0
0.5
(7) 0.0
0
formed by nonlinear regression analysis using the program INPLOT (GraphPad, Inc., San Diego, CA).
1
2
3
l/[Fru-l,68P]
4
5
4
B
RESULTS The Reverse Reaction: Initial Velocity,Product Inhibition, and Mixed Alternate Substrate Studies. Because the kinetic behavior of BsPFK in the forward direction is more complex than in the reverse direction, studies on the latter are presented first. The turnover number k,, of the reverse reaction (3.2 s-l) was about 35-fold lower than that of the forward reaction (112 s-l). This compare well to the 40-fold lower value for the reverse reaction of EcPFK relative to its forward reaction (Hellinga & Evans, 1987). Initial velocity studies were performed on the reverse reaction by varying Fru- 1,6BP concentration while keeping MgADP concentration constant at different fixed levels, and vice versa. Double-reciprocal plots were constructed for both sets of data. The families of lines obtained for both (Figure 1) intersect to the left of the ordinate, a result which denotes a sequential kinetic mechanism. The initialvelocity data were fit to the equation for a rapid-equilibrium random or ordered bireactant mechanism (eq 3), assuming that A is MgADP and B is Fru- 1,6BP. The following parameters were obtained by performing nonlinear regression analysis of the initial velocity data (Cleland, 1979) using eq (3): kcat= 3.2 f 0.2 s-l, Ka = 0.1 1 f 0.01 mM, Kb = 0.44 f 0.05 mM, Ki, = 0.33 f 0.04 mM, and Kib = 0.19 f 0.08 mM. Product inhibition studies were also performed on the reverse reaction to further distinguish whether its mechanism is random or ordered. For these studies, the concentration of one substrate (Fru-l,6BP or MgADP) was varied while the other was kept constant at a fixed level in the absence or presence of different amounts of one of the two products (Fru6P or MgATP). Four sets of product inhibition data were obtained. The data from each inhibition study were fit to the equation for either linear competitive (eq 5 ) , linear noncompetitive (eq 6), or linear uncompetitive (eq 7) inhibition. Table 1 lists the pertinent kinetic parameters obtained and the patterns of lines observed in double-reciprocal plots. The results are consistent with a rapid-equilibrium random mechanism in the reverse direction. The noncompetitive product inhibition by Fru-6P with respect to MgADP indicates the formation of a dead-end E-MgADP-Fru-6P complex. However, the equations for both competitive and noncompetitive inhibition fit equally well to the data for inhibition by MgATP with respect to Fru-l,6BP, a result which indicates that the dead-end E-MgATP-Fru- 1,6BP complex probably does not form readily, if it forms at all. Additional evidence that the mechanism in the reverse direction is random was obtained from studies using GDP, which can function as an alternate substrate for the reverse reaction. When GDP is added to a PFK assay in which
6
(mM-')
t
I
- 3 m
%S e u x P ,
e r
p1 I
2
:5 \ Is)
2
1
0'
0
"
10
"
20
"
30
l/[MgADP]
"
40
'
I
50
(mt4-l)
FIGURE1: Initial velocity patterns for the reverse reaction. (A) Plot of the reciprocal of initial velocity uersus the reciprocal of Fru-l,6BP concentration at different fixed levels of MgADP: (A)0.05 mM, (H), 0.10 mM, (0)0.20 mM, and ( 0 )0.5 mM MgADP. (B) Plot of the reciprocal of initial velocity versus the reciprocal of MgADP concentration at different fixed levels of Fru-l,6BP: (0)0.20 mM, (A)0.50 mM, (0)1.0 mM, and ( 0 )2.0 mM. The lines were drawn using linear equations containing parameters generated from direct curve-fitting to eq 3. Initial velocities (units per microgram) were measured using the Fru-6P-coupled assay. Table 1: Product Inhibition Patterns for the Reverse Reaction [ Fru- 1,6BP] [MgADP] (mM) (mM) inhibitor
pattern'
Ki, (pM) Kii (pM)
varied 0.91 MgATP* NC(orC) 5 2 f 4 676f47 MgATP C 0.50 varied 1 4 f 1 na Fru-6F NC 0.50 varied 7 f 2 55f11 Fru-6P C varied 0.91 6 f 0.5 na a NC = noncompetitive: C = Competitive inhibition. When MRATP was the product inhibitor,. the Fru-6P-coupled assay was used. k h e n Fru-6P was the product inhibitor, the ATP-coupled assay was used. na, not applicable.
NADP+ reduction is coupled to ATP via hexokinase (see Materials and Methods section), which will not accept GTP as a substrate, the GTP produced in the reverse PFK reaction is not detected and "inhibition" results. The curvature or linearity of double-reciprocal plots for initial velocity studies in the presence of GDP can provide evidence as to whether the mechanism is ordered, with ADP (GDP) binding first (curved plots), or random (linear plots). As expected, the plots obtained showed that GDP is a competitive "inhibitor" with respect to MgADPand a noncompetitive "inhibitor" with respect to Fru-l,6BP, indicating that ADP (or GDP) can bind first to the enzyme in either an orderedor a random mechanism. More significantly, the plots were linear. This result is consistent with a random mechanism, since nonlinear inhibition would have been observed had the mechanism been ordered.
Biochemistry, Vol. 33, No. 11, 1994 3427
B. stearothermophilus PFK Kinetics
A similar result has been reported for rabbit muscle PFK (Hanson et ai., 1973). Altogether, the results of both the product inhibition and the mixed alternate substrate.studies in the reverse direction are consistent with a rapid-equilibrium random mechanism that includes the formation of an abortive E-MgADP-Fru- 1,6P complex. The Forward Reaction: Fructose 6-PhosphateSaturation ofBsPFK. Valdezet ai. (1989) showed that Fru-6Psaturation of BsPFK is hyperbolic. Our findings confirm this result and further indicate that simple Michaelis-Menten kinetics cannot fully explain the saturation behavior of BsPFK. As shown in Figure 2A (dashed curve), in the presence of varied Fru-6P concentrations that are less than about one-third the fixed MgATP concentration, the Fru-6P saturation data fit the Michaelis-Menten equation well (R2= 0.994). A K,(Fru6P) of 0.03 mM is obtained under these conditions. However, when the [Fru-6P]/[MgATP] ratio is greater than about 1/3 (Figure 2A, upper curve), a flattening is observed. Thus, when analyzed over the entire 0.01-1.0 mM range, the saturation data fit the Michaelis-Menten equation less well ( R 2 = 0.988). In the presence of high levels of MgATP (greater than 30Km) Fru-6P saturation curves do not flatten within the normal 0.01-1 .O mM Fru-6P concentration range. Rather, saturation follows Michaelis-Menten kinetics (Figure 2A, lower curve) throughout the entire range. Inhibition by MgATP is apparent, and this inhibition appears competitive with respect to Fru-6P (Figure 2A, inset). Indeed, when Fru6P saturation data collected in the presence of fixed levels of MgATP equal to 0.5 mM, 5.0 mM, and 10.0 mM are analyzed according to the method of Cleland (1979), the data fit the equation for linear competitive inhibition (eq 5 ) well, yielding a linear slope replot and a Ki, of 5.2 f 0.5 mM. A Hill coefficient of 1.05 f 0.05 is obtained for each of the three saturation curves. Thus, MgATP inhibition is not associated with cooperative Fru-6P binding as it is for EcPFK, where Hill coefficients increase from 1.4 to 3.6 in the presence of fixed MgATP concentrations ranging from very low levels to above-saturating levels (Johnson & Reinhart, 1992). The flattening effect observed in the Fru-6P saturation curve when the [Fru-6P]/[MgATP] ratio is greater than about 1/3 is seen in double-reciprocal plots as a leveling-off of the slope near the 1/u axis. This leveling-off effect is evident in Figure 2B, which shows initial velocity plots with respect to variable Fru-6P concentration at different fixed levels of MgATP near its K,value. These results indicate that MgATPconcentration is rate-limiting when the [MgATP]/[Fru-6P] ratio is less than 2 or 3. The effect is more apparent in Figure 2C, which shows that the reaction rate remains steady and independent of Fru-6P concentration between 0.033 and 5.0 mM in the presence of fixed MgATP concentrations at or below its K, value (0.1 mM). Overall, these studies suggest that MgATP binds to BsPFK under non-rapid-equilibrium conditions and that the rate of catalysis is faster than the rate of MgATP binding. [The deviation from linearity at low Fru-6P concentration in the presence of 0.1 and 0.2 mM MgATP in Figure 2B is most likely the result of substrate binding via a kinetically disfavored pathway (see Discussion section).] The linear regions of the plots in Figure 2B can be fit to the initial velocity equation for a sequential bireactant mechanism in rapid equilibrium (eq 3), assuming A is MgATP and B is Fru-6P. When this is done, the following kinetic parameters are obtained: k,,, = 112 f 7 s-l, K, = 40 f 27 pM, Kb = 2 f 18 pM, and Kib = 68 f 43 pM. The doublereciprocal lines generated from curve-fitting to eq 3 intersect
.25
AI
5
tl 7
" i
0.4 0.6 0.8 [Fru-6P] (mM)
0.2
1.0
25
1.2
61
0
20
0
40 60 80 l/[Fru-6P] (mM-')
100
120
50
40 30
CI
t
20 10 0
0
10 20 l/[Fru-6P] (mM-')
30
FIGURE 2: (A) Fru-6Psaturationcurves. Dependenceof thevelocity of the BsPFK-catalyzed reaction on Fru-6P concentration in the presence of ( 0 )0.5 mM or (0) 10.0 mM MgATP is shown. The upper curve was generated by fitting the data to eq 4, whereas the lower curve was generated by curve-fitting to eq 1. The dashed line was generated by fitting the data between 0.01 and 0.2 mM FrudP in the upper curve to eq 1. (Inset) Double-reciprocal plots of the
same data. The lines were drawn using linear equations containing parameters generated from fitting the data between 0.01 and 0.2 mM Fru-6P to eq 1. Initial velocities (units per microgram) were measured using the Fru-l,6BP-coupled assay. (B) Initial velocity plot. Reciprocal of initial velocity versus reciprocal of Fru-6P concentration between 0.01 and 0.20 m M Fru-6P in the presence of (A) 0.05, (A) 0.10, and (0) 0.20 mM MgATP. The lines were drawn using linear equations containing parameters generated by direct fitting of a selected range of data points to eq 3. (C) Initial velocity plot. Reciprocal of initial velocity uersus reciprocal of Fru6P concentration between 0.033 and 5.0 mM Fru-6P in the presence of different fixed low levels of MgATP ( 0 )0.04 mM, (*) 0.05 mM, (0)0.067 mM, and ( 0 )0.10 mM. Lines were generated by linear regression analysis.
to the left of the ordinate and on the abscissa, a result consistent with a sequential mechanism in the forward direction.
3428 Biochemistry, Vol. 33, No. 11, 1994
Byrnes et al.
.251
h
1.0
I
.20
01
\
$ .15 C
v
6
3
1
0.4
5 l m 0.2
e
-0
2
4
6
8
10
0.0' 0
12
"
2
"
[MSATPI (mM)
FIGURE 3: Dependence of the BsPFK-catalyzed reaction rate on MgATP concentration. Fru-6P concentration was kept fixed at ( 0 ) 0.05 mM or (0)0.2 mM. Both curves were generated by fitting the data to eq 4. The thick dashed line was generated by fitting the data between 0.05 and 1.O mM in the upper curve to eq 1. (Inset) Doublereciprocal plots of the same data. MgATP Saturation and Substrate Inhibition. The Michaelis-Menten equation fits the MgATP saturation data well (Rz= 0.996) in the presence of saturating Fru-6P as long as the varied MgATP concentration is below 10-fold the fixed Fru-6P concentration (Figure 3, thick dashed line). A Km(MgATP) of 0.1 mM is obtained under these conditions. However, inhibition becomes apparent at high MgATP concentrations (Figure 3, both curves), and this inhibition is more pronounced at the lower Fru-6P concentration (lower curve). Furthermore, as the fixed level of Fru-6P is decreased, the inhibition becomes apparent at a lower MgATP concentration. The saturation curves tend to rise to a maximum and then decrease to a plateau. The inhibition by high levels of MgATP results in nonlinear double-reciprocal plots with respect to variable MgATP concentration. As shown in Figure 3 (inset), the doublereciprocal plots pass through a minimum and then bend upward as they approach the l / v axis. This initial velocity pattern is indicative of substrate inhibition (Segel, 1975a) by high levels of MgATP. Because of their upward curvature, the lines do not intersect to the left of the ordinate. Thus, the initial velocity data with respect to MgATP do not fit the equation for a sequential bireactant mechanism in rapid equilibrium (eq 3). MgATP Saturation of the GV212 BsPFK Mutant. It is conceivable that the inhibition exhibited by MgATP results from the binding of MgATP in a site other than the active site, e.g., the effector site. In order to address this possibility, the ability of MgATP to inhibit a mutant BsPFK enzyme that has the glycine at position 212 replaced with a valine (GV2 12) was investigated. Glycine 212 is located at the hinge of the 8H loop, which has been proposed to be important in the BsPFK allosteric transition (Schirmer & Evans, 1990). Residue 2 12 is situated along the border of the effector binding cleft near the site where the purine ring of a bound ADP or GDP molecule would be located. We have found that PEP inhibition of the GV212 mutant cannot be reversed by GDP or ADP and that neither ADP nor GDP bind well to the effector site (Zhu et al., manuscript in preparation). Presumably, binding of MgATP in the effector site would likewise be disrupted. Figure 4 shows that the ability of MgATP to inhibit the enzyme is unaffected by the mutation (the curves appear nearly identical). This observation suggests that the inhibition is not due to the binding of MgATP in the effector site.
4
"
"
6
8
"
10
"
12
[MSATPI (mM)
FIGURE 4: MgATP inhibition of the mutant GV212 BsPFK enzyme. Initial velocity was measured versus MgATP concentration for ( 0 ) mutant and (0)wild-type BsPFK in the presence of 0.05 mM Fru6P. Both curves were generated by curve-fitting to eq 4. Table 2: Alternative Nucleoside Triphosphate Substrates NTP
Km(NTP)' (mM)
5% of k,,(ATP)
K , ( F ~ U - ~ P(pM) )~
ATP 0.07 f 0.01 100 33f5 GTP 0.18 f 0.01 73 28 f 4 CTP 2.6 f 0.2 98 146 f 19 UTP 2.8 f 0.2 93 68 f 7 Fru-6P concentration was kept fixed at 1.0 mM (saturating) for thesedeterminations. NTPconcentration was kept fixed at SK,,,(NTP). In each assay, Mg2+ was present in excess of the NTP concentration.
Alternative Nucleoside Triphosphate ( N T P ) Substrates. In addition to ATP, the nucleoside triphosphates GTP, UTP, and CTP can all serve as phosphate donors in the BsPFKcatalyzed reaction. All four NTPs display saturation profiles similar to those in Figure 3, which tend to rise to a maximum and then drop to a plateau at high MgNTP concentrations. Surprisingly, although the four NTPs have different Michaelis constants (Table 2), they all give the same relative velocity (v/Vmax)at a concentration of 10 mM (not shown). Thus, they inhibit BsPFK to essentially the same extent when present at high (10 mM) concentration. Product and Dead-End Inhibition Studies of the Forward Reaction. Product inhibition studies as well as dead-end inhibition studies using the nonreactive ATP analogs 0 , ~ methyleneadenosine 5'-triphosphate (AMPPCP) and 5'adenylyl imidodiphosphate (AMPPNP) and the nonreactive Fru-6P analog arabinose 5-phosphate (Ara-SP), were performed on the forward reaction in order to further delineate the kinetic mechanism. As with the reverse reaction, primary double-reciprocal plots were constructed to determine the nature of the inhibition, and the saturation data were fit directly to equations for linear competitive, linear noncompetitive, and linear uncompetitive inhibition. The patterns obtained in the plots and the parameters obtained from curve-fitting are shown in Table 3. The product and dead-end inhibition results for the forward reaction are somewhat more difficult to interpret than are the product inhibition results for the reverse reaction. The doublereciprocal plots for inhibition by both Fru- 1,6BP and Ara-5P with respect to MgATP (up to 1 mM) yield nearly parallel yet divergent lines (Table 3). In both of these inhibition studies, when the MgATP concentration is extended to 10 mM, the plots bend upward as they approach the l l v axis, indicating substrate inhibition at high (>2 mM) MgATP concentration. Because of the curvature, it is not possible from these plots alone to establish whether the inhibition is
Biochemistry, Vol. 33, No. 11, 1994 3429
B. stearothermophilus PFK Kinetics 30
Table 3: Product and Dead-End Inhibition Patterns for the Forward Reaction [Fru-6P] [MgATP] (mM) (mM) 0.2 varied 1.o
varied 0.87 varied
(A) Product Inhibition Patterns product Kb (mM) inhibitor pattern'
Kii (mM) MgADP C 0.503 f 0.006 na 3.92 f 0.05 MgADPb NC 22 f 2 1412 Fru-1,6BPC UC (NC)I na
(B) Dead-End Inhibition Patterns* [Fru-6P] [MgATP] dead-end (mM) (mM) inhibitor pattern Ki. (mM) ~~
Kii (mM)
~
0.20
varied 0.20 varied varied 0.20
varied 0.87 varied 0.87 0.87 varied
A
AMPPCP AMPPCP AMPPNP AMPPNP Ara-5P Ara-5P
C NC C (or NC) NC C UC ( N O d
1.12f 0.04 10.2 j : 0.9 0.05 f 0.02 0.29 f 0.09 1.2 f 0.2 na
na 9.8 f 0.8 1.56 f 0.81 0.40 f 0.12 na 6.2 f 0.5
x
e
0 10 $
=.
0 -6
-4
-2
0 2 4 6 l/[MgATP] (mM-')
8
10
12
'NC = noncompetitive, C = competitive, and UC = uncompetitive. The Fru- 1,6BP-coupled assay was used for these studies. When Fru1,6BP was the product inhibitor, the ADP-coupled assay system was used. Fru-l,6BPinhibitionwith respect to Fru-6Pcould not be determined since the phosphoenolpyruvate required for the assay is an allosteric inhibitor of BsPFK at low Fru-6P concentration. Parallel lines were observed in the double-reciprocalplots. See the text for further discussion. The pattern in parentheses is that expected in the absence of substrate inhibition. na, not applicable. uncompetitive (parallel lines) or noncompetitive (convergent lines). An uncompetitive inhibition pattern would be expected in these studies if the kinetic mechanism were ordered with MgATP binding first, whereas a noncompetitive pattern would be expected if the mechanism were random. However, the ordered mechanism under consideration here is incompatible with the competitive nature of the inhibition by MgATP with respect to Fru-6P seen in Figure 2A. Competitive substrate inhibition by MgATP would be possible for an ordered mechanism only when Fru-6P is the first substrate to bind to the enzyme; it is not possible for an ordered mechanism in which MgATP binds first to the enzyme (Segel, 197Sa). On the other hand, competitive substrate inhibition by MgATP with respect to Fru-6P is consistent with a random mechanism. In this case, inhibition can occur in several ways, including the two being proposed here: (1) abortive binding of MgATP in the Fru-6P site and (2) reaction flux through a kinetically disfavored substrate binding pathway (see below). Thus, taking the substrate inhibition into account, the most likely interpretation of the product and dead-end inhibition results is that Fru-6P and MgATP bind to the enzyme in random order. An examination of the AMPPCP and AMPPNP inhibition patterns (Table 3) sheds light on the mechanism by which ATP inhibits BsPFK. Although AMPPCP inhibition with respect to ATP is purely competitive, AMPPNP inhibition with respect to ATP, though mostly competitive, nevertheless has some noncompetitive character. In the latter case, the lines in the double-reciprocal plot (Figure SA) clearly converge just to the left of the l/v axis. The simplest interpretation of this result is that AMPPNP (and presumably ATP) binds abortively in the Fru-6P site, forming a dead-end E-MgATPAMPPNP [or E-(MgATP)z] complex. Further evidence for the formation of this complex is seen in the double-reciprocal plot for AMPPNP inhibition with respect to Fru-6P (Figure SB), which gives an essentially noncompetitive (mixed-type) pattern composed of lines that interest progressively closer to the l/v axis as AMPPNP concentration is increased. This effect is not seen for AMPPCP. AMPPCP apparently does not bind abortively as does AMPPNP, most likely because the orientation of its y-phosphate group is quite different from
-40
-20
0 20 40 l/[Fru-6P] (mM-')
60
80
FIGURE5 : A M P P N P inhibition patterns. (A) Plot of the reciprocal of initial velocity versus the reciprocal of MgATP concentration in the presence of 0.2 m M Fru-6P and ( 0 )0 mM, (0)0.1 mM, (0) 0.2 mM, or (A)0.3 m M AMPPNP. (Inset) Slope replot. (B) Plot of the reciprocal of initial velocity versus the reciprocal of Fru-6P concentration in the presence of 0.87 m M MgATP and ( 0 )0 mM, (0) 0.2 mM, (0) 0.5 mM, or (A) 0.81 m M AMPPNP. (Inset) Intercept replot. Initial velocities (units per microgram) were measured using the Fru- 1,6BP-coupled assay.
that of AMPPNP, which is similar to that of ATP (Yount et al., 1971; Larsen et ai., 1969). These results suggest that ATP binds in the Fru-6P site via its y-phosphate group. Thus, abortive binding at least partially explains the inhibition seen at high MgATP concentration.
DISCUSSION The intersecting patterns of lines obtained in the doublereciprocal initial velocity plots for thereverse reaction indicate that the kinetic mechanism is sequential rather than pingpong. Further, the patterns of lines obtained in the reverse reaction product inhibition (Table 1) and mixed alternate substrate (GDP) studies indicate that the kinetic mechanism of the reverse reaction is rapid equilibrium random. The results of product inhibition and dead-end inhibition studies of the forward reaction (Table 3) are also consistent with a random mechanism, after taking into account the effect of substrate inhibition by MgATP. The binding of MgATP to the enzyme was found to be rate-limiting when its concentration is low. Thus, the kinetic mechanism of BsPFK in the forward direction can best be described as sequential random with binding of MgATP being rate-limiting, i.e., the rate of its binding is lower than the catalytic rate. The kinetic mechanism in the forward direction is depicted in Scheme 1, where E is free enzyme, A and B are substrates MgATP and Fru-6P, respectively, E-A and E-B are their binary complexes, and
3430 Biochemistry, Vol. 33, No. 11, 1994 Scheme 1 high [ A I
low 181
n E-A k9
E
E-products
E-A-8 ki o
E-B
W high 181 low [AI
E-A-B is the reactive, ternary complex. The constants klklo are rate constants for the steps indicated. It is assumed that the binding of MgATP is slow relative to the catalytic step, Le., kl, k6
